HiC input from TADbit about cloops HOT 3 CLOSED

On Tue, Oct 15, 2019 at 8:54 AM bontus ***@***.***> wrote: Dear Yaqiang, I am interested in using cLoops on our HiC data, which has been processed using TADbit and binless. Unfortunately, neither of these programs produces a BEDPE file, so I am a bit at a loss in terms of feeding this data to cLoops. TADbit does create its own BED-llike .TSV format files with paired-end mapped reads, which in turn are used by binless, so I was wondering whether these could be re-formatted and used by cLoops. I uploaded a sample file for testing purposes, see https://github.com/bontus/cLoops_TADbit/blob/master/FPR_5_both_map_chr10_sample.tsv File format description (edited based on https://github.com/3DGenomes/binless) TADbit .tsv: tab-separated text file containing paired-end mapped reads for a single experiment. All lines starting with # will be discarded. The order of the reads is unimportant. It has the following columns: id: ID of the mapped pair (alphanumeric without spaces) chr1: chromosome of read 1 (alphanumeric without spaces) pos1: position of the 5' end of read 1 (integer) strand1: whether the read maps on the forward (1) or reverse (0) direction length1: how many bases were mapped (integer) re.up1: upstream restriction site closest to pos1 (integer <= pos1) re.dn1: downstream restriction site closest to pos1 (integer > pos1) chr2: chromosome of read 2 (alphanumeric without spaces) pos2: position of the 5' end of read 2 (integer) strand2: whether the read maps on the forward (1) or reverse (0) direction length2: how many bases were mapped (integer) re.up2: upstream restriction site closest to pos2 (integer <= pos1) re.dn2: downstream restriction site closest to pos1 (integer > pos1) As I have not worked with HiC-Pro, I do not know whether its BEDPE file provides information about the positions of restriction sites or not. In case the BEDPE files for cLoops solely contain information based on the id, chr1/2, pos1/2 and strand1/2 information, one could convert the TADbit TSVs to BEDPE using AWK: TSVFILE=FPR_5_both_map_chr10_sample.tsv; grep -v "[#~]" $TSVFILE | head -n 5 \ | awk -v OFS='\t' '{print $4; if ($4 == 1 && $10 == 1) {print $2,$3,$3+$5,$8,$9,$9+$11,$1,".\t.\t+\t+";} if ($4 == 1 && $10 == 0) {print $2,$3,$3+$5,$8,$9-$11,$9,$1,".\t.\t+\t-";} if ($4 == 0 && $10 == 1) {print $2,$3-$5,$3,$8,$9,$9+$11,$1,".\t.\t-\t+";} if ($4 == 0 && $10 == 0) {print $2,$3-$5,$3,$8,$9-$11,$9,$1,".\t.\t-\t-";} }' One thing to note is TADbit's mapping, as reads can be split at the ligation site if they do not map entirely, which makes their files quite complex. An example can be seen here (after converting their TSV to SAM): D00733:158:CA67GANXX:1:1211:12600:68498#1/6 0 chr10 305327 33 1M = 430210 41 * * TC:i:1 E1:i:304175 E2:i:305356 E3:i:429917 E4:i:430247 D00733:158:CA67GANXX:1:1211:12600:68498#4/6 24 chr10 305327 33 1M = 305359 -33 * * TC:i:1 E1:i:304175 E2:i:305356 E3:i:305356 E4:i:305474 D00733:158:CA67GANXX:1:1211:12600:68498#5/6 16 chr10 305327 33 1M = 430250 -41 * * TC:i:1 E1:i:304175 E2:i:305356 E3:i:430247 E4:i:430302 D00733:158:CA67GANXX:1:1211:12600:68498~2~#2/6 0 chr10 305359 33 0M = 430210 41 * * TC:i:2 E1:i:305356 E2:i:305474 E3:i:429917 E4:i:430247 D00733:158:CA67GANXX:1:1211:12600:68498~2~#6/6 0 chr10 305359 33 0M = 430250 -41 * * TC:i:2 E1:i:305356 E2:i:305474 E3:i:430247 E4:i:430302 D00733:158:CA67GANXX:1:1211:12600:68498#4/6 24 chr10 305359 33 0M = 305327 33 * * TC:i:1 E3:i:305356 E4:i:305474 E1:i:304175 E2:i:305356 D00733:158:CA67GANXX:1:1211:12600:68498#1/6 0 chr10 430210 41 1M = 305327 33 * * TC:i:1 E3:i:429917 E4:i:430247 E1:i:304175 E2:i:305356 D00733:158:CA67GANXX:1:1211:12600:68498~2~#2/6 0 chr10 430210 41 1M = 305359 -33 * * TC:i:2 E3:i:429917 E4:i:430247 E1:i:305356 E2:i:305474 D00733:158:CA67GANXX:1:1211:12600:68498#3/6 24 chr10 430210 41 1M = 430250 -41 * * TC:i:1 E1:i:429917 E2:i:430247 E3:i:430247 E4:i:430302 D00733:158:CA67GANXX:1:1211:12600:68498~2~#6/6 0 chr10 430250 41 0M = 305359 -33 * * TC:i:2 E3:i:430247 E4:i:430302 E1:i:305356 E2:i:305474 D00733:158:CA67GANXX:1:1211:12600:68498#5/6 16 chr10 430250 41 0M = 305327 33 * * TC:i:1 E3:i:430247 E4:i:430302 E1:i:304175 E2:i:305356 D00733:158:CA67GANXX:1:1211:12600:68498#3/6 24 chr10 430250 41 0M = 430210 41 * * TC:i:1 E3:i:430247 E4:i:430302 E1:i:429917 E2:i:430247 Let me know what you think, it would be great if we can solve this issue and make cLoops work with TADbit too. Kind regards — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#15?email_source=notifications&email_token=AAOPQKOMV6HPCO3I5QJXFH3QOW4Q7A5CNFSM4JA4FTD2YY3PNVWWK3TUL52HS4DFUVEXG43VMWVGG33NNVSW45C7NFSM4HR3U46A>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAOPQKIG4E72IKGOPQDOT7LQOW4Q7ANCNFSM4JA4FTDQ> .

On Mon, Oct 21, 2019 at 3:16 AM bontus ***@***.***> wrote: Dear Yaqiang, Thanks for the advice, I am currently running cLoops with the BEDPE files from TADbit even though the split reads make up quite a high percentage (20%+) at times, so I might have to repeat this later on. So far, everything seems to work fine, but just to make sure: I split the TSV files by chromosome and thus am running each chromosome separately, is that OK for cLoops? Best — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#15?email_source=notifications&email_token=AAOPQKJAHKK22IR5UWERA2LQPVJL3A5CNFSM4JA4FTD2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEBZJXWI#issuecomment-544381913>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAOPQKLPTP75IJVVYVPNRH3QPVJL3ANCNFSM4JA4FTDQ> .